mauricio carabar n , adri an de la garza and oth on m. moreno … · 2015-04-07 · global...

GLOBAL LIQUIDITY AND CORPORATE FINANCING IN MEXICO

Mauricio Carabarın ∗, Adrian de la Garza† and Othon M. Moreno‡

Banco de MexicoThis draft: May 7 2015

Abstract. In the aftermath of the 2008 crisis, corporations in emerging market economiesincreased their issuance of debt securities in international financial markets under signifi-cantly better conditions. The undergoing process of normalization of monetary conditions,particularly in the U.S., poses potentially significant risks for these firms, especially as U.S.interest rates increase and exchange rates in emerging economies depreciate. This paper hastwo main objectives: (1) to describe the recent evolution of corporate financing in Mexicoand quantify the potential adverse effects on foreign corporate debt servicing to interest rateand exchange rate shocks; and (2) to assess how changes in global liquidity may affect boththe ability of large Mexican corporates to tap international financial markets and the accessof large and small and medium enterprises (SMEs) to domestic credit. The analysis uses anovel data set that combines credit-level data of the universe of all loans and lines of creditgranted by banks and non-banks to non-financial private firms in the country, the universeof all bond placements by these same companies, and balance sheet information of bond is-suing firms between 2003 and 2014. Our results suggest that the increase in global liquidityin the years following the Great Crisis significantly reduced international borrowing costsfor large debt-issuing firms in Mexico, which in turn explains the rise in foreign debt place-ments in recent years. We illustrate that, as large firms have turned away from domesticfinancing sources, commercial banks in Mexico have been able to funnel more resources toSMEs–a situation that may potentially reverse once international credit conditions becomemore restrictive and large firms return to domestic lending markets.

JEL classification: F32, F34, E44, G32

The contents of this article are the authors’ responsibility and do not necessarily reflect the views ofBanco de Mexico. The authors would like to thank seminar participants at the Second Federal Reserve Bankof New York-Banco de Mexico Workshop and the 6th BIS CCA Research Conference for helpful commentsand suggestions. We are grateful to Hector Reyes and Ana Laura Garcıa for their outstanding researchassistance. All remaining errors are ours.

∗E-mail: [email protected]. Av. 5 de Mayo #18, Centro 06059, Mexico, D.F. Phone: +52555237 3631.

†E-mail: [email protected]. Av. 5 de Mayo #18, Centro 06059, Mexico, D.F. Phone: +52555237 2412.

‡E-mail: [email protected]. Av. 5 de Mayo #18, Centro 06059, Mexico, D.F. Phone: +52555237 2476.

1

2 GLOBAL LIQUIDITY AND CORPORATE FINANCING IN MEXICO

1. Introduction

The years after the global financial crisis of 2008-09 have been characterized by extremely low

policy interest rates in advanced economies and abundant liquidity in international markets. Under

these conditions, risk premia declined and international corporations, particularly from emerging

economies (EMEs), expanded their placements of securities overseas, thereby satisfying asset man-

agers’ and global investors’ appetite for higher returns. This “second phase of global liquidity,”as

dubbed by Shin (2013), has been characterized by investors’ search for yield (Turner 2014) and has

been a cause of major concern for both policymakers and academics because of its potential impli-

cations for cross-border financial stability. In particular, overall portfolio flows in EMEs may have

become more vulnerable to global financial conditions, especially since it has been shown that capi-

tal flows from institutional investors into EMEs tend to be relatively stable during normal times but

pull back more strongly and persistently under stress (IMF 2014). In addition, since international

debt placements by EMEs are typically in hard currencies, the rapid increase in EMEs corporate

bond issuance may have led to currency mismatches, which could increase their vulnerability to

external macroeconomic shocks (Shin 2013).

In broad terms, the Mexican case is no different from that of other EMEs. The ample liquid-

ity in international financial markets after the 2008-09 crisis was associated with more favourable

financing conditions abroad for Mexican corporations. Not only were Mexican firms able to secure

foreign-currency financing at lower interest rates than before, but they also managed to obtain

larger amounts at longer maturities. The average maturity of bonds issued by Mexican firms

overseas increased from around 10 years in 2004 to about 16 years in 2014. The dollar amount

per issue went from an average of about 0.5 billion to more than 1.0 billion in the same period.

Perhaps more importantly, this improvement in financing conditions did not occur in domestic

debt markets–at least not to the same extent–, which favoured the search for foreign financing my

Mexican corporations. Between 2010 and 2014, the gross medium-term bond issuance abroad by

Mexican non-financial corporations averaged 13 USD billion per year, which is roughly four times

the annual average observed between 2006 and 2009.

Moreover, this surge in bond issuance overseas has implied a shift in the financing structure of

Mexican corporations. In particular, given that only the country’s largest firms have access to

international debt markets, the rise in debt placements overseas seems to have freed up financial

resources from domestic markets, which banks have been able to funnel towards firms without

access to bond markets abroad, especially small and medium enterprises (SMEs). Indeed, while

commercial bank credit to large issuing companies contracted after the 2008-09 crisis, domestic

bank credit to SMEs has grown significantly and has been the most dynamic component of banks’

commercial loan portfolios.

The main objective of this paper is to better understand how changes in global financial con-

ditions have impacted firm financing in Mexico, particularly in terms of the volume and prices of

GLOBAL LIQUIDITY AND CORPORATE FINANCING IN MEXICO 3

debt issues and bank loans. To this end, we break up the larger task at hand into three smaller

ones. First, we describe how financing to non-financial corporations in Mexico has evolved in the

last decade, focusing on the expansion of private bond issuance overseas and the terms at which

these bonds have been offered. Second, we quantify how changing global conditions during this

“second phase of global liquidity”may have affected the financing of Mexican firms in international

debt markets, particularly with respect to costs—proxied by the primary spread between an inter-

national bond and a domestic one of similar characteristics—and market activity measured by the

probability of observing a debt placement at any point in time. Third, we show that as Mexican

corporates increased their activity in international debt markets, they reduced their liabilities with

domestic banks, and estimate the extent to which this reduced bank lending to large companies may

have resulted in more credit to SMEs. In light of the current volatility in financial markets and the

expected changes in the monetary stance of advanced economies and global financial conditions in

the near future, these calculations allow us to assess both how vulnerable large Mexican corporates

may be to such changes and by how much bank lending to SMEs may decrease once international

financial markets dry up and large bond-issuing companies return to the domestic credit market.

In order to investigate these questions, we use information from original regulatory reports, in-

dividual bond prospectuses, and private data vendors to carefully assemble a proprietary data set

that combines credit-level data—including the universe of loans and lines of credit granted between

2003Q4 and 2014Q1 to non-financial private firms by banks and regulated non-banks domestically,

as well as all private bonds issued either domestically or internationally during this period—with

balance sheet information for issuing firms in Mexico. We classify firms as either large issuing firms,

large non-issuing firms or SMEs using information contained in regulatory reports on sales, number

of employees, and credit take-up. It is important to emphasize that issuing firms are examined

at the parent-company level, which implies that, for each parent company, all individual company

data were consolidated across subsidiaries.

Our results indicate that changes in global conditions in recent years have decreased the cost

of issuing debt externally relative to that paid by the same firm in the domestic market. This in

turn has increased the probability of a Mexican corporate issuing abroad. In particular, we estimate

that a 100 bp reduction in the external-internal debt issuing spread increases the probability of

issuing abroad by about 11%. In addition, our results suggest that, as Mexican corporates sought

foreign financing and decreased their credit take-up in the domestic loan market, SMEs benefited

from an increase in bank credit: a decrease of one MXN in new loans to large issuing firms increases

new loans to SMEs by 0.20 MXN. Combining these results, we estimate that a 100bp external-

internal debt issuing spread increase would reduce new bank loans to SMEs by about 4 billion

MXN, about 11% for the last observation in our sample. Although we perceive this to be a sizable

impact on bank credit to SMEs—with potential consequences to real sector activity—, it is pos-

sible that this crowding-out effect be mitigated, particularly if large issuing firms seek alternative

financing sources other than domestic bank loans, such as global bank loans, domestic debt issues


or trade credit. In addition, even if large issuing firms take up large amounts from the domestic

credit market, commercial banks may continue granting credit to SMEs given the significant push

that the recently-approved Financial Reform gives to joint loan provision between commercial and

development banks, and given low delinquency ratios and the relatively good credit history that

SMEs have built these last few years.

The rest of the paper is organized as follows. In Section 2, we present a brief review of the

literature. Section 3 describes the data collection process, provides some descriptive statistics and

describes some recent trends in corporate financing in Mexico, presenting some stylized facts with

regards to the recent surge of bond issuance in international markets by Mexican companies. Sec-

tion 4 discusses our estimation strategy and describes our main results regarding the effect of global

conditions on Mexican firms’ financing in international bond markets, particularly with respect to

costs and market activity. In Section 5, we present our findings about the substitution between

credit to debt-issuing firms and loans to SMEs. Finally, Section 6 concludes.

2. Review of the Literature

The concept of global liquidity has become increasingly popular in economic debate while re-

maining notoriously difficult to define. Most recent literature relates global liquidity to the ease

of financing conditions observed in global financial markets (BIS 2011). Empirical work produced

before the global financial crisis usually defined global liquidity in terms of the growth rate of mone-

tary aggregates from advanced economies, using either narrow or broad money measures (Baks and

Kramer 1999; D’Agostino and Surico 2009). Nevertheless, as financial integration has increased in

recent years–as evidenced by the presence of a highly integrated global financial cycle (Rey 2013)

and a larger amount of funds flowing from G4 economies to the rest of the world (Cerutti et al

2014)–, the issue of finding measures of global liquidity that are more adequate for investigating its

implications on cross-border real and financial linkages has garnered much attention.

In the aftermath of the 2008 crisis, the literature seemed to have shifted away from money-based

global liquidity measures towards the use of different types of indicators that attempt to proxy

liquidity conditions in international markets depending on the object of study. One such approach

has relied on quantity or volume indicators akin to monetary aggregates, typically employed in

studies that investigate potential financial sector vulnerabilities due to excess capital flows (Borio

et al. 2013; Bruno and Shin 2014; Cesa-Bianchi et al 2015). Another approach has been to combine

quantity and price variables, which helps in better gauging the ease of financing in global markets

(Domanski et al 2011; Chen et al 2012). Indeed, some have argued that global liquidity conditions

cannot be summarized by a single indicator (Eickmeier et al 2013) and that it is important to

differentiate between official and private liquidity–the latter having the stronger complementarity

with liquidity in domestic markets (Landau 2013). A large proportion of the recent empirical lit-

erature on global liquidity thus utilizes a quantity measure, such as cross—border credit, together

with indicators that relate to the willingness of private investors to provide funding, such as yield


spreads or risk appetite proxies like the VIX index (Domanski et al 2011; Chen et al 2012; Cerutti

2014).

Irrespective of the way it is measured, there is much agreement in the literature that global liquidity

has important effects on real and financial variables on both advanced and emerging economies,

particularly serving as “a vehicle for the numerous interactions and spillovers between domestic

monetary and financial policies”(Landau 2013). For example, Canova (2005) use a VAR model to

show that U.S. monetary shocks produce significant fluctuations in Latin American macroeconomic

variables, being the interest rate channel an important amplifier of these shocks on output, inflation

and nominal exchange rates in the region. Ruffer and Stracca (2006) employ a similar approach to

conclude that a global monetary policy shock has a positive effect on prices and output in the euro

area and Japan. More recently, Bayoumi and Bui (2011) show by means of event studies that the

U.S. monetary stimulus packages, particularly QE1, were associated with large reduction in emerg-

ing market yields and currency appreciations, and generally supported foreign activity. McCauley

et al (2015) use panel estimation techniques to illustrate how accommodative U.S. monetary policy

and cheap leverage have promoted credit growth globally. Finally, Azis and Shin (2013) studied the

effect of more ample global liquidity conditions in the years following the great financial crisis on

Asian economies. In particular, they find that rising non-core bank liabilities played an important

role in the transmission of global liquidity shocks to emerging Asia, and that the growing share of

foreign ownership of equity and securities in the region makes these markets more susceptible to

volatility and capital outflows.

This paper makes three key contributions to the literature on global liquidity and its impact on

domestic macrofinancial conditions. First, while most empirical and theoretical studies analyse the

potential spillover effects of excess capital flows via the banking sector, this analysis emphasizes

the role of international corporate debt markets. This is particularly relevant in the aftermath

of the 2008 crisis given the generalized boom in bond issuance in global markets by non-financial

firms in emerging economies. Second, our paper uses a comprehensive data set of the universe

of bond placements and individual commercial loans provided by the domestic banking sector to

study how global financial conditions changed the structure of credit and debt markets. To the

best of our knowledge, this is the first paper that uses microdata to this end. Third, in contrast to

most of the literature that relies on cross-country studies, panel regressions, and VAR models, our

data set allow us to employ a microeconometric approach to study the impact of global liquidity in

one particular case—that of the Mexican corporate sector. Although the Mexican case may differ

from that of other emerging economies, we believe that the results derived from this analysis may

provide some useful insights to academics and policymakers interested in understanding the po-

tential spillover effects that lax financial conditions globally may have had on developing economies.

Finally, our research also contributes to the literature regarding the liability structure of firms


and the ownership composition of debt portfolios. On the topic of firms’ liability structure, no-

table contributions include Colla et al (2012) and Morellec et al (2013), who present evidence of

higher diversification across multiple debt types for large rated firms, as well as for firms with

more growth options and higher bargaining power. These results are in line with Rauh and Sufi

(2010) who document debt heterogeneity in firms’ balance sheet debt, and important adjustments

in the underlying composition of corporate debt, even though no significant one year changes in

total debt are observed. Moreover, in what concerns what drives firms to issue bonds in foreign

markets, as opposed to domestic markets, Gozzi et al (2012) explore the use of international debt

market by firms using a database with information on new bond issues at a transactional level

including bonds from several countries. In particular, they find that international bond markets

complement the domestic markets, thus providing different financial services, other than issuing in

foreign currencies, not fulfilled in domestic markets. As shown in the next section, these findings

are corroborated by our description of the characteristics of domestic and foreign debt placements

by Mexican corporates.

3. Data description and recent evolution of corporate financing in Mexico

3.1. Data description.

In order to investigate the effect of global conditions on firm financing decisions in Mexico,

we gather a comprehensive dataset with credit-level and firm-level information on Mexican non-

financial private corporations for the period of 2003 Q4 to 2014 Q1. We complement these data

with a number of both global and domestic macroeconomic aggregates, including those aimed at

measuring the availability of financial resources in international markets. The remainder of this

section describes the sources and the main features of the data.

Credit-level data (individual bonds and loans)

We obtain data on the universe of corporate bonds issued either domestically or abroad by Mexi-

can non-financial private corporations between 2003 Q4 to 2014 Q1. Information for bonds issued

in international markets is retrieved from Bloomberg, which provides the main characteristics on

each issuance (e.g. issuer, issue and maturity dates, amount issued, yield at issue, credit rating,

currency, type of rate, periodicity of coupon payments). We fill in some missing data from firms

individual bonds prospectuses, and add a few other indicators from this source, such as the use

of proceeds from the debt placement. Tables A1 and A2 summarize the data. In total, there are

226 international bond issues by 53 non-financial Mexican firms during the sample period, most of

which are denominated in USD (87.6% of the number of bond issuances) and of maturities greater

than a year (99% of outstanding debt). The majority of these bonds were issued during the period

associated to the “second phase of global liquidity”—about two-thirds of bonds were issued after

the collapse of Lehmann in 2008 Q3. In terms of their characteristics, bonds issued internationally

after the crisis observed lower yields, larger amounts, and slightly longer maturities on average.

Although these are simple averages that do not necessarily reflect the greater heterogeneity of firms


participating in international bond markets, they are broadly consistent with what would be ex-

pected in an environment of ample liquidity and looser financial conditions.

For debt issued domestically, we collect individual bond-level data in a similar way. The main

sources of information are Indeval and Valmer, which provide financial data services on debt issued

by Mexican public and private entities. We fill in missing data using other sources, particularly

Bloomberg and individual bond prospectuses. To make these bonds comparable with those issued

abroad, we focus only on debt placements with maturities greater than a year. Tables A3 and A4

show some descriptive statistics. Our database consists of 316 bonds issued in domestic markets

during the sample period by 82 different firms. As expected, most of these bonds (87.3%) are

denominated in MXN, while 11% of these bonds are denominated in inflation-indexed units, UDIs.

Only four bonds in our sample were denominated in USD. In contrast to the marked increase in in-

ternational bond issuance after the 2008 crisis, domestic bond placements rose only slightly—about

53% of domestic private debt issues occurred after September 2008. In terms of their character-

istics, yields on domestic bonds decreased after the crisis, and the average amount per bond was

somewhat larger. However, the average maturity decreased slightly in the post-2008 period.

Finally, we also obtain information on the universe of loans provided by regulated financial inter-

mediaries in Mexico to non-financial private corporations using Banco de Mexico’s credit registry.

The data come from monthly regulatory reports (series R-04 C) administered by Banco de Mx-

ico and the National Banking and Securities Commission (CNBV for its Spanish acronym), which

provide transaction-level information on loans and lines of credit granted by regulated financial

intermediaries (banks and non-banks) to firms and individuals with entrepreneurial activity (or

personas fısicas con actividad empresarial in Spanish). We restrict our sample to non-financial

firms, which are identified by their RFC—a unique ID used for tax collection purposes. The sam-

ple period ranges from December 2003 to April 2014. For any given month, we observe individual

loans and lines of credit for each firm, as well as the main characteristics of the credit, including

interest rate, maturity, original amount, remaining balance, among others. As expected, bank

credit is much more widespread among Mexican firms than domestic or international bonds, at

least in terms of the sheer number of loans and lines of credit granted nationwide. In January 2014,

for example, there were 584,135 individual loans and lines of credit granted to 120,459 unique RFCs.

Using the above data, we classify firms as large companies or SMEs. Large companies are either

debt issuers (with participation in debt markets, either domestic or foreign) or non-debt-issuing

companies that meet at least one of the following conditions: (i) received a loan larger than 100

MXN million at any time during our sample period; reported having more than 250 employees; or

reported annual sales of more than 250 MXN million. In order to identify loans granted to large

issuing firms, it is important to acknowledge that many of these firms obtain financing through

their subsidiaries, which have their own RFCs. Thus, we collect information on the names of the

subsidiaries of all listed companies in Mexico, which they are required to disclose on their quarterly


financial reports. We used these names to find the RFC of each subsidiary and consolidate all data

across subsidiaries at the parent-company level.

Firm-level data

Balance sheet and general financial data on non-financial private corporations in Mexico come

mainly from financial reports of listed companies on the Mexican Stock Exchange (Bolsa Mexi-

cana de Valores, or BMV). These reports are published quarterly and their financial information

is electronically available in Banco de Mexico’s data system from 2007 onwards. We complement

this information to expand the sample period to begin in 2003 Q4 by manually collecting financial

data contained in individual reports retrieved directly from the BMV, as well as by using Econo-

matica, a private data provider. These efforts allow us to put together financial information on

139 different firms, whose dollar-volume of debt placements represents close to 90% of the total

dollar amount issued during the sample period. Individual firm characteristics include total assets,

which we employ as a measure of the size of the company. We use assets and sales denominated

in foreign currency to measure the “natural hedges”of these companies’ foreign liabilities. We also

collect information on firms’ profitability, liquidity, and short-term financing needs, such as ROA

and short-term assets and liabilities. We summarize these data in Table A5.

Figure 1 summarizes some of the information described above. During our sample period, 182 non-

financial private firms in Mexico obtained financing through either international or domestic debt

placements, or through the stock exchange. Out of these, 53 firms issued bonds in international

markets and 82 were active in domestic debt markets. We have balance sheet information on all

firms listed on the stock exchange. For non-listed firms the only data available is that related to

their activity in domestic credit markets, and their individual bond placements if they issued at

all.

3.2. Recent trends in corporate financing in Mexico.

The financing structure of firms in Mexico has undergone substantial changes after the global

financial crisis, highlighted by a significant increase in financing through external bond issuance

(Figure 2). In September 2008, external bond issuance represented 14% of total firm financing to

non-financial corporations in Mexico; by 2014 this figure has roughly doubled, which implies that

financing through external bond issuance has increased threefold in real terms during the period.

The growth rates of domestic financial sources have been positive but relatively much lower, while

external credit to these firms has decreased, congruent with the process of deleveraging by banks

in advanced economies following the crisis.

During the period that this shift in the liability structure of Mexican firms occurred, domestic

commercial bank credit to large firms with debt issuance – either abroad or domestically – began

to display a downward trend (Figure 3). It is important to note that this is only observed for issuing

firms: credit to large non-issuing firms has continued to grow in recent years, albeit at a slower pace


than credit to SMEs. From September 2008 to July 2009, the period when international financial

markets were most disrupted by the crisis, the trends described above were strikingly different:

during these months the amount of outstanding loans to large non-issuing firms stalled, loans to

SMEs decreased and loans to issuing firms surged.

Private bond placements in international markets by Mexican corporations increased significantly

in recent years. Annual gross bond issuance by private non-financial corporations in international

markets averaged 13 USD billion in 2010-2014, which contrasts with the average of 4 USD billion

recorded in 2002-2009. Net issuance averaged 10.1 USD billion in 2010-2014, up from 0.3 USD

billion in 2002-2009 (Figure 4). This surge in bond issuance has been mirrored by an increase in

the number of firms issuing in international markets. For example, in 2013 there were 21 private

non-financial corporations that issued bonds abroad, seven of them for the first time. In 2003, this

figure was much smaller, with only five firms issuing bonds abroad.

Greater access to international financial markets has implied greater diversification along several

characteristics for the set of Mexican firms issuing abroad. Firms from a wider variety of sectors

have been able to tap into these markets (Figure 5a). In 2003, around 70% of gross bond issuance

in international markets was from firms in telecommunication services. In contrast, this figure was

close to 30% in 2014, as sectors that previously had a very limited participation in these markets

have increased their issuance, such as materials and consumer discretionary and services sectors.

In addition, bonds with a wider variety of grades have been issued recently, congruent with the

entrance of new firms in the market and a greater appetite for risk and search for yield by interna-

tional investors (Figure 5b). Issuing terms for Mexican firms in international bond markets became

more favourable in the years following the 2008-09 crisis. Interest rates have become lower (Figure

6), while the average size and maturity of the issuances have increased. As Figure 6 shows, this

trend is true even when limiting the set of issuances to those made by firms with activity in both

domestic and international markets.

4. The effect of global liquidity on external bond issuance

The aim of this section is to quantify the potential impact of the “second phase of global liq-

uidity”on corporate bond issuance by Mexican firms in international markets. In particular, we

are interested in identifying how changes in financing conditions abroad associated with the ample

liquidity in financial markets may have affected the probability that a firm in Mexico issue an

international bond:

(1) Prob(Issuei,∗t = 1) = f(Spreadi,t , θ) + εi,t

The dependent variable is the probability that firm i issues a bond in international markets in

period t. This probability is a function of, among other things, the cost of placing this bond abroad


relative to what it would cost the firm to place a bond with similar characteristics (particularly

amount and maturity) in the domestic market. This cost is referred to as a spread or yield differ-

ential between the yield of an international bond and that of a domestic issue, as faced by firm i

at any given point in time. Defined this way, an increase in this spread would imply rising relative

costs of issuing debt abroad, which should in principle have a negative effect on a firm’s decision

to issue an international bond.

There are both theoretical and empirical challenges in estimating a model of the form we de-

scribed. First, there is no clear consensus in the literature as to what exactly drives corporate debt

financing in international markets (Gozzi et al 2012). The presence of frictions in financial markets

(at home or abroad) may affect firms’ financing decisions, particularly since debt securities and

loans may be characterized by different price and non-price attributes (Black and Munro 2010).

Non-price attributes could be related to several factors such as: (i) depth of market, as firms are

more likely to match their financing needs with lenders’ preferences in deeper markets, thus making

it easier to cumulatively raise more funds with a single or fewer issuances (Modigliani and Sutch

1966; Vayanos and Vila 2007); (ii) currency hedging, as issuances abroad may provide a useful nat-

ural hedge to their assets and cash flows denominated in foreign currency (Munro and Wooldridge,

2009); and (iii) funding diversification. Price attributes could be relevant because firms may also

issue abroad simply because the coupon rate of an international debt placement may become less

costly relative to that of a domestic bond.1

Besides these more theoretical issues, analyzing the particular case of Mexican firms presents

additional empirical challenges because bond market participation (either domestically or inter-

nationally) is relatively low, as shown in Section 3. The rather small number of individual bond

issues, as well as the fact that companies do not typically issue bonds in both markets around the

same time, hampers our ability to compare international and domestic debt financing costs for a

given firm. For example, during the period from 2003 Q4 to 2014 Q1, there were only 25 times

when a single firm issued bonds in both domestic and external markets in the same quarter. These

25 observations correspond to only eight different corporations. The interest rate differential for

a given period across markets may only be calculated in a straightforward manner for these few

observations. This problem is particularly relevant since the global liquidity effect may operate

mainly through the price channel (Shin 2013). Also, the price component is observed only when a

firm issues a bond, which immediately raises a well-known selection bias problem that should be

addressed.

To mitigate these concerns, we construct a counterfactual individual issuer-level spread time series

as a function of global financial conditions, domestic macroeconomic conditions, as well as indi-

vidual firm and bond characteristics. To do this, we first estimate the primary market coupon

1More generally, this question is related to a literature that links capital flows in EMEs to the effect ofboth push and pull factors (Calvo, Leiderman and Reinhart 1993, 1996; Schularick and Taylor 2012).


rate that each firm would pay both internationally and abroad for a bond of a given amount and

maturity. We then take the difference of these yield rates to construct the spread.

To be more precise, we first estimate two separate equations of the following form:

(2) y{D,∗}i,t = f(Global/DomesticV arst, BondChars

{D,∗}i,t , F irmCharsi,t) + ε

{D,∗}i,t

where the dependent variable is the observed primary market coupon rate for a bond issuance of

firm i in period t, either in domestic or international markets (D, *). We restrict our data to

bond issues in MXN and UDIs for the domestic market (99% of domestic issues) and in USD for

international markets (88% of international issues) in order to facilitate the comparison of interest

rates across currencies—as explained in more detail below. In terms of the particular controls use

in this estimation, we include global financial conditions, domestic macroeconomic indicators, and

individual firm and bond characteristics.

In terms of the proxies that reflect global financial conditions, we aim to capture the effect of

global liquidity on primary market coupon rates through price and quantity indicators as discussed

in Section 2. In particular, we consider as volume measure the total credit to non-residents in USD

and euros 2, obtained from the BIS Global Liquidity selected indicators. We also include the VIX

index as a price-based indicator of global liquidity reflecting investor’s risk appetite, and thus prox-

ying investors’ willingness to provide funding (Eickmeier et al 2013). In addition, we incorporate

the U.S. corporate BBB/BAA-Treasury 10-year spread in order to capture financing conditions

particular to debt markets, especially since flucutations in this particular spread should closely

mimic financing conditions faced by emerging economies’ corporates in international markets as

they approximately belong in the same asset class.

Regarding domestic conditions, we control for domestic activity using the year-on-year real GDP

growth rate in Mexico, and include the 28-day interbank interest rate (TIIE 28) as a proxy for do-

mestic financing conditions. Individual firm characteristics include a binary variable that indicates

whether the firm has ever issued an investment grade bond in external financial markets, as well

as lagged values of the firm’s total assets and return over assets (ROA). The latter two attempt

to measure firm size and profitability, which should have an impact on financing costs. Finally, we

control for maturity-at-issue and original amount of the bond.

The results presented in Table 1 correspond to the model where the international bond inter-

est rate is the dependent variable. Estimation (1) includes only bond characteristics. Estimation

(2) incorporates our global liquidity measures. Estimation (3) adds domestic conditions, and finally

estimation (4) includes firm characteristics. We find that the amount issued has a negative effect

2McCauley et al (2015), for example, use the same measure, but restricted to USD.


on the interest rate of international bond issuances, while maturity is not statistically significant in

any of our estimations. We find that cross-border credit is associated with a negative sign and is

always significant, which suggests that the abundance of liquidity in financial markets has reduced

the cost of issuing abroad to Mexican firms. In contrast, the U.S. corporate BBB/BAA-Treasury

10 year spread is associated positively to the yield, which suggests that, as risk appetite in debt

markets decreases, financing costs for Mexican firms increase. The VIX index is not statistically

significant in any of the estimations, although this is likely due to the high correlation with the

U.S. corporate spread (94% correlation). Finally, we also find that our variables concerning firm

characteristics all have the expected signs: firm profitability in the previous period seems to have a

negative effect on the interest rate paid in external bond issuances, while being a firm that has is-

sued bonds with investment grade in international markets also has a negative and significant effect.

The results from the model where the domestic bond interest rate is the dependent variable are

reported in Table 2. These results are ordered in a similar fashion to the previous table, as we add

different sets of variables to the model. Global conditions are found to have qualitatively similar

effects than in the model for external bond issuances, although neither the U.S. corporate spread

nor the VIX are significant in estimations (3) and (4). The TIIE 28 is positive and statistically

significant, which, unsurprisingly, implies that as domestic financing conditions become tighter, the

interest rates paid by firms in domestic debt issuances also rises.

With these results in hand, we obtain fitted values from the models that incorporate bond and

firm characteristics, as well as global and domestic conditions. We use these fitted values to com-

pute the interest rate that a given firm would have paid at each point in time and in each market

(D, ∗) for a bond with standard characteristics (500 million USD and 1 year maturity):

(3) {y∗, yDMXN}i,t

Then, in order to compare prices across markets, we transform the rates of bonds issued domesti-

cally by adjusting for expected depreciation:

(4) yDUSD =S

F(1 + yDMXN )− 1

Where yDMXN is the primary market coupon rate of domestic bonds, yDUSD is the primary market

coupon rate of bonds issued abroad, S is the spot exchange rate and F is a 1 year forward exchange

rate.

Finally, we calculate the spread at each point in time and for each firm as:


(5) Spread = y∗ − yDUSD

We plot the counterfactual spread we calculated in Figure 7, while Figure 8 and 9 present the

individual external and internal coupon rates for selected firms, respectively. The obtained spread

displayed a substantial increase at the height of the global financial crisis, when conditions in in-

ternational financial markets were most severely disrupted. Afterwards, there is a clear downward

trend in the spread, which implies that the relative cost of issuing abroad for Mexican firms has

been decreasing during the aftermath of the global financial crisis. This is congruent with the

increase of bond issuances in international markets by Mexican firms and the search for yield by

global investors during the period.

Having estimated this (counterfactual) spread for each individual firm, we then estimate how

cheaper financing conditions abroad may affect international bond issuance through a reduced-form

binary choice model as in equation (1) above. Our results are summarized in Table 3. Estimation

(1) controls only for whether a firm has issued in external markets before. Estimation (2) adds our

measure of global cross-border credit in hard currencies. Estimation (3) adds additional domestic

controls, such as domestic bank lending capacity.3 Estimation (4) adds other firm characteristics,

such as the ratio of short-term assets to short-term liabilities in order to control for firms’ financ-

ing needs. Using this last estimation, evaluating at means, we find that a 100bp reduction in the

foreign-domestic debt issuing spread increases the probability of issuing abroad by 10.5%. We also

find that having past experience in international debt markets is significant and has a positive im-

pact on the probability of issuing abroad across all estimations. Finally, our measure of cross-border

credit is not significant in any estimation, which suggests that most of its effect arises through prices.

5. The effect of global liquidity on domestic credit

In this section, we investigate how changes in global financial conditions have affected the do-

mestic credit market in Mexico. In particular, we are interested in examining whether the increase

in bond issuance in international debt markets by large debt issuing firms has reduced their par-

ticipation in domestic banks loan portfolio, thus freeing up resources that could then be allocated

for financing SMEs. Our motivation for estimating whether changes in global financial conditions

have made an impact on loans to SMEs stems from the stylized facts presented in Section 3 and

3 We use total obligations net of repo and derivatives operations since the banks’ balance sheets inMexico include transactional accounts used to record these types of operations that could overestimate theavailable loanable funds. In particular, we use the following definition for lending capacity::

LendCapn = TOg +max{0, derl − dera}+max{0, repol − repoa};where LendCapn is lending capacity, TOg is gross total obligations excluding derivatives and repo operations,derl and repol are the derivatives and repo operations on the liability side of the balance sheet, and similarlyfor the asset side.


our regression results in Section 4. First, we showed evidence of a change in the liability structure

of debt issuing firms—some of the largest in the country—consisting of a rapid increase in external

debt issuance coupled with a downward trend in outstanding volume of loans granted to these

firms. In Figure 10 we plot the share of external debt on total financing of non-financial corpora-

tions in Mexico that are either listed on the stock exchange or have participated in debt markets

throughout our sample period. This share has exhibited an increasing trend since September 2009,

growing from representing 39.6% of total financing to 64.3%. At the same time, loans to other

firms without participation in debt markets have displayed positive growth rates, particularly loans

to SMEs. Finally, our estimations in Section 4 showed that the increase in global liquidity has

implied a decrease in the relative price of issuing abroad that a given firm faces, which in turn has

increased the probability of issuing international markets.

Finding evidence of this crowding-in effect of loans to SMEs in Mexico is relevant for several

reasons. SMEs represent a large share of economic activity in Mexico: they are estimated to ac-

count for 72% of total employment in the formal sector and their economic activity is estimated

to represent 34% of GDP as of 2013 (Plan Nacional de Desarrollo, 2013). These figures are largely

consistent with the findings of Ayyagari, Beck and Demirguc-Kunt (2007), who report that SMEs

account for around 60% of manufacturing employment in a large sample of developed and devel-

oping countries. However, SMEs typically experience greater difficulty in obtaining resources from

the financial sector (Beck, Demirguc-Kunt, Laeven, and Maksimovic, 2006). The main theoretical

argument for explaining this phenomenon is that SMEs are often more informationally opaque than

larger firms, which in turn aggravates the agency problem inherent in corporate lending relation-

ships, and even more so for larger banks that tend to rely more on “hard”quantitative information

in their lending screening process (Berger and Udell 1998, Berger and Black 2011). Also, SMEs

are more likely to face financing constraints, since they usually have a smaller variety of financing

sources available for them.

Both of these mechanisms seem to be at work in Mexico, since SMEs tend to face higher interest

rates in the domestic commercial credit market than larger firms and also display slightly higher

delinquency rates. Additionally, SMEs in Mexico do not currently have access to debt markets.

According to the Credit Market Survey conducted by Banco de Mexico, around 23% of firms with

less than 100 employees received credit from commercial banks in December 2014, and none issued

debt. Most of these firms received financing through trade credit (around 65%). Finally, it is also

important to investigate the determinants of the volume of commercial bank credit to SMEs in

Mexico because of the relative scarcity of research devoted to the subject.4

In order to investigate whether there is indeed a crowding in effect of loans granted by commercial

banks to large debt-issuing companies on loans provided to SMEs, we estimate an equation of the

4 Lecuona (2009) and Fenton and Padilla (2012) provide general overviews of the topic.


following form:

(6) SMEt = f(Issuert, γ) + ut

where Issuer is a measure of loans granted by commercial banks in Mexico to debt issuing firms

and SME represents loans granted to SMEs. However, the main challenge in estimating the model

described above consists in overcoming a potential endogeneity problem. Since this is a reduced

form model, it is not difficult to argue that there may be other variables not included in the model

that are relevant in determining the amount of loans granted to both SMEs and debt issuing firms.

We attempt to address this issue by using an instrumental variable approach, informed by the

results in the previous section. In particular, we attempt to use variables that reflect the condi-

tions in international financial markets as instruments for the credit granted to issuing firms. In

particular, we use again the VIX index as a measure of risk appetite in financial markets, since it

reflects the availability of funding in international markets.5 This is potentially a valid instrument

since domestic credit to SMEs should not be directly affected by this variable since they themselves

do not have access to financing in international debt markets, and because we are also controling

for credit supply channels, like net bank obligations, and real demand channels, such as economic

growth in the country. At the same time, we have already showed how global variables affect the

participation in international markets of debt issuing firms, which in turn affects their demand for

loans in the domestic market. Furthermore, our estimations only use new loans granted in a given

period, which ameliorates concerns that the persistence inherent in using outstanding credit could

drive our results. Descriptive statistics on new loans for each type of firm is summarized in Table

A6. Finally, we control for the availability of resources in the domestic credit market by controlling

for our measure net bank obligations.

Thus, we estimate the following model:

SMEt =β1Issuert + β2LendCapt+Xγ + ut(7a)

Issuert = θ1V IXt + Zµ+ εt(7b)

where Issuer are new bank loans granted to debt issuing firms in a given period; SME are new

loans granted to SMEs in that same period; LendCap is our measure of domestic banks’ lending

capacity; X are additional variables meant to control for other macroeconomic aggregates that may

affect domestic credit market conditions, such as bank’s average funding costs and the economy’s

growth rate (measured by the IGAE index); and V IX represents the VIX index. The data for this

estimation has a monthly frequency and our sample ranges from December 2013 to April 2014.

5We run the same estimations using the US corporate spread as an additional instrument that capturesthe availability of funding in international financial markets. The results remain qualitatively unchanged.


Our results are summarized in Table 4. The first two regressions were estimated by OLS and

the coefficients associated with new loans to issuing firms are not significant. Once we implement

our instrumental variable strategy—in regressions (3) and (4)—, we find that this coefficient is

negative and statistically significant across all models. We interpret this result as evidence of a

crowding-in effect of loans to SMEs when debt issuing firms turn away from the domestic credit

market. Our measure of net bank obligations is positive and also statistically significant across

all models, which themselves differ as we use more macroeonomic variables as controls. The in-

strument we use is the monthly average of the VIX index, a standard measure of the market’s

expectation of stock market volatility. While perhaps the more natural variable to use as an instru-

ment would have been the firm-level interest rate spread we calculated before, we did not do it for

two reasons. First, we would have lost information for firms for which we were unable to build a

spread (e.g. firms with missing balance sheet data). Most importantly, the interest rate spread we

built had a quarterly frequency. As expected, the VIX has a significant and positive effect on new

loans granted to debt issuing firms in the first stage of our IV regressions, which suggests that, as

conditions tighten in international debt markets, large firms turn to the domestic credit market for

financing. Our preferred estimation is regression (4) in Table 4, which has additional controls for

macroeconomic aggregates. Using the coefficients associated with new loans to debt issuing firms,

we find that a decrease of 1 billion MXN in new loans to large issuing firms increases new loans to

SMEs by around 200 million MXN.

Finally, we are interested in using our results for estimating the potential impact of a tightening of

conditions in international financial markets on the volume of loans to SMEs. This is particularly

relevant in light of the expectation that there will be an increase in the policity interest rates of

some advanced economies–mainly the United States—in the coming months. If this increase in

the policy rates occurs and markets tighten, then our results suggest that Mexican debt-issuing

firms would face higher interest rates in debt markets abroad relative to the domestic market,

which would generate incentives for them to find financing in the domestic credit market. This

increase in the demand for loans by debt issuing firms could then crowd-out SMEs from the do-

mestic credit market, which could potentially hamper their productive activity as they have fewer

financial sources at their disposal.

In order to perform this calculation, we need to make some additional assumptions. We first

assume that gross bond issuance in external markets decreases in the same proportion as a de-

crease in the probability of issuing abroad by any firm, as estimated in Section 5. Furthermore, we

assume that these debt issuing firms that are unable or unwilling to place bonds in international

substitute completely their financing needs for domestic bank loans. That is, the decrease in debt

issuance abroad is symetricall to their increase in loans obtained through the domestic credit mar-

ket. This would mean that their position in other debt markets –such as the external credit market,

domestic debt market or trade credit—remains unchanged. Under these assumptions, combining


our results from our preferred estimations in Table 4, we find that an increase in 100bp in the

interest rate spread would translate into a crowding-out of new loans to SMEs of approximately

4 MXN billion, approximately 11% of the last observation in our sample. In turn, an increase of

1 standard deviation in the interest rate spread—equivalent to a surge of 159 basis points—would

lead to a crowding-out of new loans to SMEs of 6.4 MXN billion (17% of the last observation in

the sample).


6. Conclusions

In this paper, we studied how the abundance of global liquidity in the aftermath of the global

financial crisis affected firm financing in Mexico. We gathered both firm-level and credit-level infor-

mation in order to study how this change in global conditions had an effect across several markets

by changing Mexican firms’ financing structure. We found that the increase in global liquidity

seems to have reduced the cost of international borrowing for large debt-issuing firms in Mexico

relative to the cost of issuing bonds domestically, which in turn may explain the significant rise in

international debt placements in recent years. We also found evidence that shows that firms with

access to international bond markets—which are Mexico’s largest firms—changed their financing

structure during the period, substituting domestic financial sources for external sources. Among

domestic sources, we found that the amount of credit channeled through domestic financial inter-

mediaries displayed a downward trend during the period. Our estimations suggest that, as large

firms have turned away from the domestic credit market, banks’ capacity to funnel domestic credit

resources to SMEs has increased, generating a crowding-in effect on these firms. This is a particu-

larly relevant finding given that this situation may be reversed once international credit conditions

become more restrictive and large firms return to domestic lending markets.

It is important to bear in mind that this crowding-out effect may be mitigated if large issuing

firms seek alternative financing sources other than domestic bank loans (e.g., international bank

credit, domestic debt issues, trade credit). Additionally, credit conditions may soon become more

restrictive in domestic markets as well. Finally, it is possible that banks continue granting credit

to SMEs even if large firms return to the market, particularly if SMEs have displayed good credit

records since their entrance.


References

[1] Avdjiev, S., McCauley, R. and McGuire, P. (2012). “Rapid credit growth and international

credit:challenges for Asia”. Bank for International Settlements, 377.

[2] Ayyagari, M., Beck, T., and Demirguc-Kunt, A., 2007. “Small and Medium Enterprises across the

Globe”. Small Business Economics 29: 415-434.

[3] Azis, I.J. and Shin, H.S. (2013) “How do global liquidity phases manifest themselves in Asia?”. Asian

Development Bank.

[4] Baks, K., and Kramer, C. (1999). “Global liquidity and asset prices: Measurement, implications, and

spillovers”. International Monetary Fund 99-168.

[5] Bayoumi, T. A. and Bui, T. (2011). “Unforeseen Events Wait Lurking: Estimating Policy Spillovers

from US to Foreign Asset Prices”. International Monetary Fund, 11.

[6] Beck, T., Demirguc-Kunt, A., Laeven, L., and Maksimovic, V. (2006). “The determinants of financing

obstacles”. Journal of International Money and Finance, 25: 932-952.

[7] Berger, A. N., and Black, L. K. (2011). “Bank size, lending technologies, and small business finance”.

Journal of Banking and Finance, 35: 724-735.

[8] Berger, A. N., and Udell, G. F. (1998). “The economics of small business finance: The roles of private

equity and debt markets in the financial growth cycle”. Journal of Banking and Finance, 22: 613-673.

[9] Black, S. and Munro, A. (2010). “Why issue bonds offshore?”. Bank for International Settlements, 52.

[10] Borio, C., McCauley, R. and McGuire, P. (2011). “Global credit and domestic credit booms”. BIS

Quarterly Review, 3: 43-57.

[11] Borio, C. and Zhu, H. (2012). “Capital regulation, risk-taking and monetary policy: a missing link in

the transmission mechanism?”. Journal of Financial Stability, 8 236-251.

[12] Calvo, G. A., Leiderman, L. and Reinhart, C. M. (1993). “Capital inflows and real exchange rate

appreciation in Latin America: the role of external factors”. Staff Papers-International Monetary Fund,

40: 108-151.

[13] Calvo, G. A., Leiderman, L. and Reinhart, C. M. (1996). “Inflows of Capital to Developing Countries

in the 1990s”. The Journal of Economic Perspectives, 10: 123-139.

[14] Canova, F. (2005). “The transmission of US shocks to Latin America”. Journal of Applied econometrics,

20: 229-251.


[15] Cerutti, E., Claessens, S., and Ratnovski, L. (2014). “Global Liquidity and Drivers of Cross-Border

Bank Flows”. International Monetary Fund 14-69.

[16] Cesa-Bianchi, A., Cespedes, L. F., and Rebucci, A. (2015). “Global liquidity, house prices and the

macroeconomy: evidence from advanced and emerging economies”. Bank of England, 522.

[17] Chen, M. S., Liu, M. P., Maechler, A. M., Marsh, C., Saksonovs, M. S., and Shin, M. H. S. (2012).

“Exploring the dynamics of global liquidity”. Andrews McMeel Publishing, 12-246.

[18] Chung, K., Lee, J. E., Loukoianova, E., Park, H., and Shin, H. S. (2014). “Global liquidity through the

lens of monetary aggregates”. International Monetary Fund, 14.

[19] Colla, P., Ippolito, F. and Li, K. (2012). “Debt specialization”. The Journal of Finance, 68: 2117-2141.

[20] Bank for International Settlement, Committee on the Global Financial System. (2011). “Global

Liquidity-Concept, Measurement and Policy Implications”.

[21] D’Agostino, A., and Surico, P. (2009). “Does global liquidity help to forecast US inflation?”. Journal

of Money, Credit and Banking, 41:479-489.

[22] Domanski, D., Fender, I. and McGuire, P. (2011). “Assessing global liquidity”. BIS Quarterly Review,

4: 57-71.

[23] Eickmeier, S., Gambacorta, L. and Hofmann, B. (2014). “Understanding global liquidity”. European

Economic Review, 68: 1-18.

[24] Fenton, R., and Padilla, R. (2012).“Financiamiento de la banca comercial a micro, pequeas y medianas

empresas en Mexico”. Serie Estudios y Perspectivas, 135: 1-66.

[25] Gobierno de Mexico. Plan Nacional de Desarrollo 2013-2018.

[26] Gozzi, J. C., Levine, R., Peria, M. S. M. and Schmukler, S. L. (2012). “How firms use domestic and

international corporate bond markets”. National Bureau of Economic Research, 17763.

[27] International Monetary Fund. (2014). “Global Financial Stability Report- Moving from Liquidity- to

Growth Driven Markets”.

[28] Landau, J. P. (2013). “Global liquidity: Public and private”. Manuscrito para la Reunin Jackson Hole

2013.

[29] Lecuona, R. (2009). “El financiamiento a las pymes en Mexico, 2000-2007: el papel de la banca de

desarrollo”. Serie Financiamiento del Desarrollo, 207: 3-50.


[30] McCauley, R. N., McGuire, P., and Sushko, V. (2015). “Global dollar credit: links to US monetary

policy and leverage”. Bank for International Settlements, 483.

[31] Modigliani, F., and Sutch, R. (1966). “Innovations in interest rate policy”, American Economic Review,

56(1/2): 178-197.

[32] Morellec, E., Valta, P., and Zhdanov, A. (2013). “Financing Investment: The Choice between Bonds

and Bank Loans”.

[33] Munro, A., and Wooldridge, P. (2009). “Motivations for swap-covered foreign currency borrowing”.

Bank for International Settlements, 2-42. Preliminary draft.

[34] Rauh, J. D. and Sufi, A. (2010). “Capital structure and debt structure”. Review of Financial Studies,

23: 4242-4280.

[35] Rey, H. (2013). “Dilemma not trilemma: the global financial cycle and monetary policy indepen-

dence”. Proceedings of the Federal Reserve Bank of Kansas City Economic Symposium at Jackson Hole.

[36] Ruffer, R. and Stracca, L. (2006). “What is global excess liquidity, and does it matter?”European

Central Bank, 696.

[37] Shin, H. S. (2013). “The second phase of global liquidity and its impact on emerging economies”. In

Keynote address at the Federal Reserve Bank of San Francisco Asia Economic Policy Conference.

[38] Schularick, M., and Taylor, A. M. (2012). “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles

and Financial Crises, 18702008”. American Economic Review, 102: 1029-1061.

[39] Turner, P. (2014). “The global long-term interest rate, financial risks and policy choices in EMEs”.

Bank for International Settlements, 441.

[40] Vayanos, D., and Vila, J. (2007). “A preferred habitat model of the term structure of interest rates”.

NBER Working paper.


Tables and Figures

Table 1. Dependent variable: International bond interest rate

(1) (2) (3) (4)Bond characteristicsMaturity -0.033 -0.041 -0.039 0.012

(0.052) (0.060) (0.058) (0.038)Amount (ext) -1.405*** -1.347*** -1.355*** -0.229

(0.227) (0.276) (0.271) (0.195)Global conditionsCross-border credit -0.346** -0.397* -0.499***

(0.134) (0.228) (0.122)US corporate spread 0.470 0.717 1.108*

(0.514) (0.590) (0.649)VIX 0.027 0.008 -0.016

(0.051) (0.057) (0.050)Domestic conditionsGDP Mex (yoy) 0.047 0.052

(0.082) (0.090)TIIE 28 -0.056 -0.059

(0.192) (0.128)Firm characteristicsInvestment grade (ext) -2.146***

(0.557)Total assets (lag) -0.236

(0.158)Profitability (lag) -0.073**

(0.033)

Observations 157 157 157 97Adj. R-squared 0.490 0.534 0.529 0.658

Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1


Table 2. Dependent variable: Domestic bond interest rate

(1) (2) (3) (4)Bond characteristicsMaturity 0.124*** 0.142*** 0.125*** 0.112***

(0.020) (0.028) (0.028) (0.029)Amount (int) -0.564*** -0.472*** -0.470*** -0.250*

(0.161) (0.115) (0.105) (0.138)Global conditionsCross-border credit -0.790*** -0.227*** -0.241**

(0.088) (0.073) (0.092)US corporate spread 1.542*** 0.571 0.504

(0.336) (0.356) (0.483)VIX -0.123*** -0.031 0.000

(0.040) (0.033) (0.040)Domestic conditionsGDP Mex (yoy) 0.011 0.041

(0.050) (0.055)TIIE 28 0.612*** 0.564***

(0.091) (0.135)Firm characteristicsInvestment grade (ext) -0.645*

(0.347)Total assets (lag) 0.019

(0.131)Profitability (lag) -0.037

(0.023)

Observations 233 233 233 146Adjusted R-squared 0.138 0.463 0.548 0.525

Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1


Table 3. Regression results: Probability of issuing abroad

(1) (2) (3) (4)Logit model

Issuance spread (fitted values) -0.107* -0.108* -0.118* -0.113*(0.058) (0.059) (0.064) (0.065)

Has issued in external market 0.788*** 0.698** 0.700** 0.688**(0.246) (0.278) (0.279) (0.279)

Cross-border credit 0.053 -0.055 -0.055(0.077) (0.271) (0.272)

Domestic conditions NO NO YES YESFirm characteristics NO NO NO YES

Observations 1,130 1,130 1,130 1,130Pseudo R-squared 0.0208 0.0215 0.0218 0.0233

Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

Table 4. Regression results: Effect of Global Liquidity on Domestic Credit

(1) (2) (3) (4)New loans to SME

OLS IVIssuer (new loan) -0.004 -0.009 -0.088*** -0.200***

(0.020) (0.027) (0.034) (0.065)Net lending capacity 46.868*** 42.260*** 47.413*** 51.422***

(2.161) (3.120) (2.200) (5.000)Bank’s avg. funding cost -0.631** 0.712

(0.315) (0.646)IGAE (Mex growth rate) -0.223* -0.641***

(0.130) (0.194)Large (new loan)

Constant Yes Yes Yes Yes

Observations 124 124 124 124F-Stat (first stage) 90.82 35.51Adj-R2 0.804 0.809 0.778 0.727

Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

We instrument Issuer (new loan) with the VIX index. VIX index is possitive and significantat p<0.01 in the first stage regressions of estimations (3) and (4). F-stat in these estimationsare 90.8 and 35.5, respectively.


Figure 1. Participation in (International/Domestic) Debt and Stock Markets byFirms in Mexico

Total 137

With external debt issuance 53

With internal debt issuance 82

Number of firms

18 firms

72 firms 24 firms

10

33

22

3

International Debt Market

Domestic Stock Market

Domestic Debt Market

Source: Banco de Mexico with information from BMV, Bloomberg and Indeval.


Figure 2. Financing to Non-Financial Corporations in Mexico.Index, 2008q3=100

0

50

100

150

200

250

300

350

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Domestic credit

Domestic bond issuance

External credit

External bond issuance

Source: Banco de Mexico.


Figure 3. Commercial Bank Credit to Non-Financial Corporations in Mexico byFirm Type1/.MXN billion

0

50

100

150

200

250

300

350

400

450

500

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Large firms, debt issuers

Large firms, non-issuers

Small and medium firms

Source: Banco de Mexico.1/ Only includes loans denominated in MXN.


Figure 4

(a) Gross Bond Issuance by MexicanCorporations in International Bond

Markets1/.USD billion

0

2

4

6

8

10

12

14

16

18

20

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Non-investment grade

Investment grade

(b) Number of Mexican CorporationsIssuing Bonds in International

Markets1/2/

7 6

8

14

11

15

3

15 14

15

18

22

16

1 1 1

4

2

4

1

34

0 1

56

0

5

10

15

20

25

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Source: Prepared by Banco de Mexico based on information from Bloomberg.1/ Includes bonds with maturity greater than one year issued in the external market by

non-financial private corporations.2/ The database has information since 1992, which means that new entrants for each year are

those that issued bonds in international markets in the corresponding period but had notdone so since 1992.The number of firms that match this criterion is represented by the white

figure inside each column.


Figure 5

(a) Gross Bond Issuance by MexicanCorporations in International

Markets by Sector.Percent

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Energy Industrials

Consumer discretionary & Services Materials

Consumer staples Telecommunication services

(b) Gross Bond Issuance by MexicanCorporations in International

Markets by Grade1/.USD billion

0

2

4

6

8

10

12

14

16

18

20

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

BB

BBB

A

B

CCC-D

NR

Source: Prepared by Banco de Mexico based on information from Bloomberg.1/ Corresponds to the grade awarded to the issuer in the corresponding foreign market based

on Bloomberg Composite.

Figure 6

(a) Interest Rates of BondsIssued Abroad by Mexican

Corporations withInvestment Grade1/.

Percent

0

1

2

3

4

5

6

7

2003q4-2008q2

2010q3-2014q1

1 to 5 yrs

5 to 10 yrs

15+ yrs

10 to 15 yrs

(b) Average Maturity ofBonds Issued by Mexican

Corporations2/.Years, 3-year moving average

0

2

4

6

8

10

12

14

16

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Domestic market

International market

(c) Average Maturity ofBonds Issued by Mexican

Corporations2/.Years, 3-year moving average

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Domestic market

International market

Source: Prepared by Banco de Mexico based on information from Bloomberg and Indeval.1/ Includes only bonds denominated in USD.

2/ Includes only corporations that issued both in the domestic and foreign market.


Figure 7. Counterfactual Spread (International Minus Domestic Interest Rate).Percentage points

0

1

2

3

4

5

6

7

8

9

10

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Investment grade issuer

Non-investment grade issuer

Source: Prepared by Banco de Mexico based on information from Bloomberg.


Figure 8. Fitted and observed external coupons.Percent

0

4

8

12

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

AMX

0

4

8

12

16

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

CEMEX

0

4

8

12

16

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

GEO

0

4

8

12

16

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

VITRO

0

4

8

12

16

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

URBI

0

4

8

12

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

BIMBO

0

4

8

12

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

FEMSA

0

4

8

12

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

ALFA

0

4

8

12

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

TELEVISA



Figure 9. Fitted and observed internal coupons.Percent

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

AMX

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

CEMEX

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

GEO

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

VITRO

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

URBI

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

BIMBO

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

FEMSA

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

ALFA

0

4

8

2004

2004

2005

2006

2006

2007

2008

2008

2009

2010

2010

2011

2012

2012

2013

2014

TELEVISA



Figure 10. Share of external bond issuance in total financing for firms in ourmain sample.

Percent

0

10

20

30

40

50

60

70

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014


Appendix

Table A1 Foreign bond issuances. Includes bonds issued in foreign markets by non-financial firms residing in Mexico. Issuances are considered investment grade when rated BBBor higher under BBG composite rating.

Full sample Pre 2008Q3 Post 2008Q3

Number of firms 53 32 40

All currenciesNumber of bond placements 226 77 149Investment grade 84 22 62Non-investment grade 142 55 87

USD onlyNumber of bond placements 198 73 125Investment grade 58 20 38Non-investment grade 140 53 87

Table A2 Foreign bond issuances: Main features. Includes bonds issued in foreignmarkets by non-financial firms residing in Mexico. Yield is the annualized primary marketcoupon rate. Amount is measured in millions of US dollars. Maturity represents maturity atissuance expressed in years.

Full sample Pre 2008Q3 Post 2008Q3Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.

All currenciesYield 7.6 3.2 8.2 2.9 7.3 3.3Amount (millions) 376.6 348.9 308.6 257.4 411.8 383.9Maturity (years) 9.7 9.1 9.2 6.6 10.0 10.1

USD onlyYield 8.0 3.1 8.1 3.0 7.9 3.2Amount (millions) 340.4 323.2 302.0 258.9 362.8 354.5Maturity (years) 8.5 6.5 8.5 5.6 8.5 6.9


Table A3 Domestic bond issuances. Includes bonds issued in the domestic market bynon-financial firms residing in Mexico. Issuances are considered investment grade when ratedBBB or higher under BBG composite rating.


Number of firms 82 43 61

All currenciesNumber of bond placements 316 146 170Investment grade 294 140 154Non-investment grade 22 6 16

MXN and UDI onlyNumber of bond placements 312 146 166Investment grade 291 140 151Non-investment grade 21 6 15

Table A4 Domestic bond issuances: Main features. Includes bonds issued in the domes-tic market by non-financial firms residing in Mexico. Yield is the annualized primary marketcoupon rate. Amount is measured in millions of US dollars.Maturity represents maturity at issuance expressed in years


Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.All currenciesYield 8.2 2.3 9.4 2.1 7.1 1.9Amount (millions) 129.0 129.7 115.0 108.3 141.0 144.8Maturity (years) 7.5 6.4 7.9 6.9 7.1 5.9

MXN and UDI onlyYield 8.2 2.3 9.4 2.1 7.2 1.9Amount (millions) 126.7 125.5 115.0 108.3 137.0 138.3Maturity (years) 7.5 6.4 7.9 6.9 7.1 6.0


Table A5 Firm characteristics. All variables measured in millions of MXN pesos, exceptROA (Return over assets) which is the ratio of net income over total assets.

Mean Std. Dev. Min Max N

Total assets 42,750.8 116,775 5.48 1,383,459.0 3,955Short-term assets 13,546.5 30,197 0.29 400,008.3 3,955Short-term liabilities 9,622.3 27,654 7.27 376,741.9 3,955ROA 0.2 0 -2.34 1.1 3,902Assets in foreign currency 6,556.3 3,6610 0.00 573,925.4 2,128Outstanding foreign debt (bond issuance only) 2,295.4 15,989 0.00 425,877.9 6,552Outstanding domestic debt (bond issuance only) 1,274.0 4,143 0.00 64,344.0 6,552

Table A6 Domestic credit by type of firm. Debt issuer are non-financial firms residing inMexico that have issued bonds in either the domestic or international debt markets throughoutthe sample period. Large non-issuer are non-financial firms residing in Mexico that have beengranted a loan of over 100 million MXN pesos at any point in time during our sample period,have 250 or more employees, or have reported sales of over 250 million MXN pesos, whileremaining inactive in debt markets. SMEs are non financial firms residing in Mexico that donot fulfill any of the previous criteria. The volumen of new loans represents loans grantedby regulated financial intermediaries residing in Mexico to private non-financial firms withresidence in Mexico, expressed in billions of MXN pesos.

Volume of new loansMean Std. Dev. Min Max

TotalDebt issuer 39.3 15.2 18.5 9 .3Large, non-issuer 52.7 17.3 26.3 99.8SME 22.7 8.0 10.3 43.1

Pre 2008Debt issuer 37.2 15.3 18.5 75.8Large, non-issuer 38.5 8.1 26.3 63.0SME 16.5 4.4 10.3 30.7

Post 2008Debt issuer 41.0 15.0 19.0 97.3Large, non-issuer 64.2 14.0 38.2 99.8SME 27.8 6.7 15.8 43.1

mauricio carabar n , adri an de la garza and oth on m. moreno … · 2015-04-07 · global...

Documents